Data from: Assessing spatio-temporal priorities for species’ recovery in broad-scale dynamic landscapes

1. As threats to biodiversity from environmental change increase, assessing priorities for mitigation action becomes increasingly important. However, there have been few attempts to schedule actions across broad spatial extents that explicitly account for dynamic ecological processes and threats. 2. We combined a dynamic occupancy model with a decision analysis framework to spatially allocate multiple recovery actions to maximize species’ probability of occupancy under threats posed by climate and land-use change. We used the koala Phascolarctos cinereus across the Australian state of New South Wales (810 000 km2) to illustrate this approach. We considered four actions implemented on a 10 × 10 km2 grid: reduce domestic dog attacks through dog control, reduce vehicle collisions through fencing highways, protect habitat through land acquisitions and restore Eucalyptus forest. We used the occupancy model to predict ecological responses to recovery actions and simulated annealing to identify spatio-temporal priorities for each action. We contrasted the results against priorities generated using a traditional static distribution model. 3. To maximize the probability of koala occupancy in 50 years’ time, with an annual budget of up to AU$20 million, investment priorities were located in the eastern part of koala's range, focusing on dog control with some investment in habitat protection and restoration. With higher budgets, investment priorities shifted towards habitat protection and restoration in the western part of the range. However, priorities based on the static distribution model, which had a lower predictive accuracy than the dynamic model, were different. Regardless of budget, priorities derived from the static model were predominantly located in the western part of koala's range, focusing on highway fencing with some investment in dog control. 4. Synthesis and applications. Our approach for integrating spatio-temporal dynamics into conservation prioritization across broad spatial extents provides a significant advance on existing approaches based on static distribution models. The finding that the inferior static model produced different priorities to the dynamic model cautions against the use of static models for conservation planning in dynamic landscapes. Additionally, the substantial changes in priorities with budget indicate that conservation planning under dynamic landscape and climate change must carefully consider priority actions and locations relative to the conservation resources available